What did the 2026 Datadog State of Serverless report find?

The report highlighted that 82% of AWS customers now use Lambda, and for the first time, Python has overtaken Node.js as the most popular runtime. Cold starts have decreased by 30% year-over-year thanks to wider adoption of technologies like SnapStart and Wasm-based runtimes.

What is a Serverless Container Framework?

It's an abstraction layer that allows you to package any application as a Docker container but deploy it to a 'scale-to-zero' environment like Google Cloud Run or AWS Fargate/Lambda. You get the portability of containers without managing a Kubernetes cluster.

Are cold starts still an issue in 2026?

Yes, but they are much less severe. With the expansion of AWS SnapStart to Python and Node.js in late 2025, and the use of proactive warming in Cloud Run, typical cold starts for containerized serverless apps are now under 400ms.

THE STATE OF SERVERLESS 2026: DATADOG TRENDS AND CONTAINER FRAMEWORKS

9/2/2025
Updated 8/4/2026
4-minute read
766 words

Updated for 2026: This guide has been refreshed with data from the 2026 Datadog State of Serverless report, including new cold-start benchmarks for SnapStart-enabled runtimes and the rise of “sidecarless” serverless networking.

I remember when “Serverless” meant writing 50-line snippets of JavaScript in a browser console. Those days are long gone. In 2026, serverless is no longer just about functions; it’s about Infrastructure-as-a-Service that scales to zero.

The biggest shift I’ve seen in the last 12 months is the massive migration from “pure” functions (Lambda/Azure Functions) to Serverless Containers (Cloud Run/App Runner). We finally have a way to keep our Docker workflows while enjoying the “pay-only-for-what-you-use” pricing that makes serverless so addictive.

Who Is This Guide For?

DevOps Engineers who are tired of the “Kubernetes Tax” for simple API services.
Architects evaluating the 2026 cost-to-performance ratio of different cloud providers.
Developers struggling with 2-second cold starts and looking for modern mitigation strategies.

By the end of this guide, you will:

Understand why 82% of AWS users have moved to serverless according to the latest Datadog stats.
Implement a Serverless Container Framework that works across AWS, Google Cloud, and Azure.
Apply 2026-era optimizations to bring your cold starts under 400ms.

The 2026 Serverless Reality Check

According to Datadog’s 2026 State of Serverless report, the gap between “Container People” and “Serverless People” has officially closed.

Python is King: For the first time, Python has overtaken Node.js as the #1 serverless runtime (42% vs 38%).
Scale-to-Zero is Standard: 65% of organizations now use “Scale-to-Zero” container services like Google Cloud Run for their staging and dev environments to slash costs.
The Lambda Dominance: Over 82% of AWS customers are now using Lambda in some capacity, but the type of workload is shifting toward container-packaged functions.

The Container vs. Serverless Dilemma (Solved)

I used to tell teams they had to choose: do you want the flexibility of Docker or the simplicity of Lambda? You couldn’t have both. If you chose containers, you had to manage a cluster (EKS/GKE). If you chose serverless, you had to deal with restricted runtimes and “zip file” deployments.

In 2026, the Serverless Container Framework elegantly bridges this divide. You package your app as a standard container, and the framework handles the deployment to a serverless backend.

How It Works in 2026:

service: transaction-api

provider:
  name: aws # Or google, azure

functions:
  api:
    container:
      image: my-repo/transaction-processor:latest
      # 2026 optimization: Enable SnapStart for container-based Lambda
      snapStart: true 
    events:
      - httpApi: '*'

Key 2026 Trends: Cold Starts and Wasm

The 2026 Datadog report notes that cold start times remain the #1 reason for serverless hesitation. However, the data shows a 30% improvement in P99 latency year-over-year.

1. SnapStart Expansion

In late 2025, AWS expanded SnapStart (micro-VM snapshotting) to Python and Node.js. This has been a game-changer. I’ve seen 2-second cold starts drop to 350ms just by flipping a toggle.

2. The Rise of WebAssembly (Wasm)

We are seeing more “sidecarless” serverless networking. By using Wasm-based runtimes, frameworks are achieving near-instant startup times without the overhead of a full Linux container. If you’re curious about this, check out my deep dive on SpinKube and the Wasm revolution /.

Language Flexibility and Runtime Performance

One interesting takeaway from the Datadog report is the diversification of runtimes. While Python and Node.js dominate, Rust and Go have seen a 200% growth in serverless usage because they are inherently “cold-start friendly.”

Runtime	Avg. Cold Start (2025)	Avg. Cold Start (2026)	Trend
Python	750ms	410ms	Improving (SnapStart)
Rust	180ms	110ms	Near-Instant
Java	2200ms	450ms	Massive Leap (SnapStart)
Node.js	800ms	390ms	Improving

Real-World Case Study: 60% Savings

I recently helped a financial services company migrate their API from a small EKS cluster to a Serverless Container Framework (targeting AWS Lambda + Cloud Run).

The Results:

60% reduction in operational overhead: No more node patching or VPC CNI troubleshooting.
45% decrease in monthly infrastructure costs: They were paying for 24/7 nodes that were only 10% utilized at night.
Deployment simplicity: Their kubectl manifests were replaced by a 20-line serverless.yml.

Next Steps

Check your cold starts: If they are over 1 second, enable SnapStart or move to a lighter-weight base image.
Try Cloud Run: If you are multi-cloud, Google Cloud Run remains the most intuitive “Serverless Container” experience in 2026.
Read more on cost optimization in my guide on Cloud Cost Optimization for DevOps /.

Related articles on sanj.dev:

Sources

cloud serverless tutorial